JPH04122844A - Inspecting apparatus for defect - Google Patents

Abstract

PURPOSE:To remove background noise and to discriminate a defect accurately without being affected by the noise by a method wherein a prescribed density range containing a background density is turned to be of a fixed density. CONSTITUTION:An analog signal (x) is converted into a digital density signal (y) of prescribed gradation in an A/D converter 32. This density signal (y) is converted in terms of density gradation in a density conversion means 34 and displayed in a TV monitor 36. For conversion of the density gradation, a table of characteristics stored in a table means 38 is used. This table is so prepared that a fixed density is provided in a fixed density width (a) on the opposite sides of a background density (c), and outside this width (a), conversion into a maximum density (255) and a minimum density (0) is made with an inclination (b). Watching an image in the TV monitor 36, an operator operates a setting means 40 so that noises in a background area may disappear. The density signal after the density conversion is compared with a prescribed threshold value in a multi-value-coding means 42 and multi-value-coded, e.g. binary-coded. When it is larger or smaller than the prescribed value, it is judged to show a defect by a defect detecting means 44 and a defect signal is outputted.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to a defect inspection device that inspects the presence or absence of defects using a density signal obtained by digitizing an image signal obtained by scanning an inspection object. .

(Technical Background of the Invention) A defect inspection 414 is known in which the surface of a steel plate, plastic film, paper, etc. is optically scanned to detect scratches on the surface or internal defects.

Conventionally, an image signal obtained by scanning an inspection object is compared with a preset threshold value, and if the image signal is larger or smaller than this threshold value, it is determined that there is a defect.

However, in this method, if the image signal contains a noise component, this noise component may be determined to be a defect, and there is a problem in that the defect detection accuracy decreases.

(Objective of the Invention) The present invention has been made in view of the above circumstances, and it is an object of the present invention to eliminate the influence of noise and improve defect detection accuracy when the image signal obtained by scanning the inspection object contains noise. The purpose of the present invention is to provide a defect inspection device that can improve the quality of defects.

(Structure of the Invention) According to the present invention, it is an object of the present invention to provide a defect inspection apparatus that detects defects in an inspection object using a density signal obtained by digitizing an image signal obtained by scanning an inspection object. table means for outputting a density gradation conversion table for converting a predetermined density range including a predetermined density range into a constant density; density conversion means for converting the density signal from the density signal using the density gradation conversion table; and a density signal after the density conversion. A defect inspection device comprising: a multi-value conversion means for converting the multi-value density signal into a multi-value signal; and a defect detection means for detecting as a defect an image portion where the multi-value density signal exceeds a predetermined density level and outputting a defect signal. ,
This is achieved by

If the density signal after density conversion is subjected to spatial filtering processing using a differential filter and then multivalued, the outline of the defect image is emphasized and the defect detection accuracy is further improved. If image smoothing processing is performed after spatial filtering processing using a differential filter, detection due to noise is further reduced and defect detection accuracy is further improved.

Further, it is even more convenient if the density signal converted into density or the multivalued density signal can be displayed on a monitor, since defects can be visually observed.

Although a single density gradation conversion table may be used here, it is desirable to be able to change settings such as the background density of the conversion table manually or automatically depending on the object to be inspected. Alternatively, a plurality of tables with different settings may be stored in advance, and the table to be used may be selected manually or automatically depending on the object to be inspected. When changing settings or selecting a table to be used automatically, the background density can be determined using a density histogram within a predetermined image area.

(Function) The digitized density signal is converted to a constant density using the density gradation conversion table, where noise included in a predetermined density range including the background density becomes the background density. Partial noise disappears. Therefore, noise disappears from the background of the image, improving defect detection accuracy. Further, even if the density signal after density conversion is subjected to spatial filtering processing using a differential filter, background noise is not emphasized, and defect detection accuracy is improved.

(Embodiment) FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention, and FIG. 2 is a diagram showing an embodiment of a density gradation conversion table and input/output density signals.

In this embodiment, the settings of the conversion table can be changed manually while viewing the image displayed on the monitor, and the image is wound from the supply roll 12 to the take-up roll 14. 16
is a laser light source. A laser beam 18 emitted from this light source 16 is guided onto a main scanning line 22 of the inspection object 10 by a rotating polygonal mirror 20, and scans the main scanning line 22 from left to right. This main scanning line 22
A light guiding rod 24 is disposed opposite to the main scanning line 22 , and the reflected light from the surface of the inspection object 10 of the laser beam 18 scanned on the main scanning line 22 enters the light guiding rod 24 . Photomultipliers 26 (26a, 26b) are attached to both ends of the light guide rod 24, and the intensity of the reflected light entering the light guide rod 24 is detected by the photomultipliers 26. Ru.

The outputs of both photomultiplier tubes 26a and 26b are added together by an adder 30 via a preamplifier (not shown) to obtain an analog image signal χ. This analog image signal χ is the inspection object 1
00 This is a gray level signal corresponding to the surface or internal properties.

This analog signal χ is converted by the A/D converter 32 into a digital density signal y of a predetermined gradation (for example, 256 gradations). This density signal y is subjected to density gradation conversion in the density conversion means 34 and displayed on the television monitor 36. Here, for the density gradation conversion, a table of characteristics stored in the table means 38, for example, shown in FIG. 2 is used. This second
In the figure, the horizontal axis indicates the density of the input density signal y,
The vertical axis indicates the output density signal Y after density conversion. These density signals y1Y have, for example, 256 gradations. In addition, this table assumes a constant density (for example, an intermediate density of 128) within a certain density width a on both sides of the background density C, and this width a
Outside of , it is converted to maximum (255) and minimum density (0) with slope b.

In this embodiment, the set values a, b, and c of this table can be changed by manual setting means 40. That is, the operator operates this setting means 40 while viewing the image on the television monitor 36 so that the noise in the background area disappears.

The density signal after density conversion is compared with a predetermined threshold value in the multi-value converting means 42, and is multi-valued, for example, binarized.
When it is larger or smaller than a predetermined threshold value, the defect detection means 44 determines that it is defective and outputs a defect signal. The results are recorded in a defect recording means 46 such as a printer along with the coordinates of the defect. This defect signal may be sent to other equipment and used for other processing. Further, a multivalued image of a certain area including a portion determined to be defective by the defect detection means 44 may be outputted to the monitor 36.

FIG. 3 is a block diagram of the second embodiment, and FIG. 4 is a density histogram. The table means 38A of this embodiment is configured to automatically set the conversion table to be used using a density histogram. That is, based on the A/D converted density signal y, the histogram means 50 obtains a density histogram of the density signal y on a fixed image area, for example, the -scanning line, as shown in FIG. From this histogram, the background density determining means 52 sets the density at which the maximum frequency is obtained as the background density C. Furthermore, the set values a and b explained in FIG. 2 can also be determined from this histogram. For example, a setting value a corresponding to the level of noise included in the background and a setting value s corresponding to the density level distribution of defects are determined from the degree of spread of the density distribution. In this case, the set value a is set from the concentration range where the frequency n0/2 is half of the maximum frequency n0, and the average slope or maximum slope between the concentrations where the frequency n0/10 is n0/2 and 1/10, for example. The slope b can be determined from

When the setting values a%b, c are determined in this way, the table creation means 54 first obtains a conversion table pattern using the setting values a, b, and then matches the setting value C corresponding to the background density level of this pattern. Set the table to the second
Translate horizontally in the figure. The density conversion means 34 converts the input density signal y using the conversion table determined in this way, obtains the density signal Y, and displays it on the television monitor 36. Alternatively, a and b may be set in advance, and a conversion table may be set by combining them with C obtained from the histogram.

This density signal Y is also subjected to spatial filtering processing by a differential filter 56 to emphasize the outline of the defect. As shown in Fig. 5, this process superimposes a spatial filter F on a 3x3 area centered on each pixel of the image to be processed, calculates the product of the corresponding pixels, and calculates the sum as the output value. shall be. This operation is performed in rask scanning order from the upper left pixel to the lower right pixel.

As a spatial filter used in this filtering process, for example, a filter having weighting coefficients as shown in FIG. 6 can be employed. This uses two filters: a filter Δ, f for row direction differentiation and a filter Δ, f for column direction differentiation, and the sum of the absolute values of the outputs of both filters or the larger value of the outputs of both filters is used. This is the final output value.

As a result of such filtering processing, the contours of the image in the chi and y directions are emphasized, and the defect detection accuracy in the subsequent multi-value processing 42 is improved. Moreover, the density conversion means 34
Because the noise in the background area is erased, the noise will not be emphasized.

FIG. 7 is a block diagram of the third embodiment. In this embodiment, a table means 38 stores density gradation conversion tables of a plurality of different patterns in a memory (table memory) in advance.
B, so that the table to be used can be manually selected by selection means 40B. That is, the operator looks at the image appearing on the television monitor 36 and manually determines a conversion table that eliminates noise from the background and makes defects most visible.

Further, in this embodiment, the density signal is multivalued through a differential filter 56 and a smoothing means 58. Smoothing means 5 here
8 is not erased by the density conversion means 34 and is not erased by the differential filter 56.
This method removes the isolated point noise that has been emphasized through differential processing to further clarify the distinction between the background and defects. That is, in this smoothing process, for example, as shown in FIG. ). If the number of pixels exceeding this density threshold (pixel number threshold) is a certain number (for example, 4) or more, the value of the pixel targeted for processing is set to 1, otherwise it is set to O. By doing so, it is converted into a binary value.

Although the smoothing means shown in FIG. 8 simultaneously performs smoothing and binarization by comparing with a predetermined threshold value, they may be performed separately. For example, the median value of the density of each pixel in the window of FIG. 8 can be used as the density of the pixel to be processed (median filter), or the average value of the density of each pixel can be used as the density of the target pixel. In this case, it is necessary to further process this by the multivalue converting means 42.

FIG. 9 is a block diagram of the fourth embodiment. In this embodiment, the conversion table shown in FIG. 7 is automatically determined. That is, this selection means 40C uses the histogram means 50 described in FIGS.
Find it in 2. The conversion table closest to the determined setting conditions is read out from the table means 38C that stores a large number of conversion tables, and density conversion is performed using this table.

The above embodiment uses a flying spot method in which the laser beam scans the surface of the inspection object, but the inspection object is irradiated with a rod-shaped light source arranged in the width direction, and the reflected light is sent to the receiver via a rotating mirror. It may be a flying image type reading method, or a type reading a line sensor or an area sensor.

The density conversion table may be stored as a two-dimensional map, but may also be stored in the form of a -order or multi-order equation. It is desirable that the quantization level of the A/D converter, the size and weighting coefficient of the differential filter, the threshold value of the smoothing means, etc. can be set freely depending on the object to be inspected.

In the above embodiment, detected defect information is recorded in a recording means such as a printer, but it is also recorded in a recording medium such as a floppy disk or a magneto-optical disk, a visual system such as a lamp or a buzzer, a contact signal, a computer, or other devices. The defect information may be converted to an appropriate format and level and then output to a signal processing device or the like.

Although the above embodiments have been described as detecting defects appearing on the surface of the object to be inspected, the present invention also includes detecting defects inside by scanning the surface. For example, internal defects in a steel plate may be detected using a magneto-optical effect.

This detects the leakage magnetic field from defects when the inspected material is magnetized with an alternating magnetic field as a change in the polarization of reflected light.

(Effects of the Invention) As described above, the present invention converts a predetermined density range including the background density into a constant density, so that noise contained in the background can be removed without being affected by the noise. It becomes possible to accurately identify defects. (Claim (1)).

Only one type of concentration conversion table may be used here, but if the inspection target changes, the settings of the conversion table may be changed manually (
[Claim (21)] It is desirable that the change can be made automatically using a histogram (Claim (3)).

In addition, multiple conversion tables may be stored in advance (claims (
41) Even if one of them is configured to be selected manually or automatically, various inspection targets can be handled (Claim +51, f6]).

Further, by subjecting the density-converted density signal to spatial filtering processing using a differential filter, the outline of the defect can be emphasized and the accuracy of defect detection can be improved (Claim (7)). Furthermore, if the density signal after the spatial filtering process is further smoothed, it is possible to remove the isolated point noise that remains in the density conversion means 34 and is emphasized by the subsequent differential processing, thereby making it possible to detect defects. Accuracy is further improved (claim (8)). Concentration converted here 4
If the density signal or the multivalued density signal is displayed on the monitor, it is convenient because the image of the defect can be viewed at the same time (Claim +91, flol).

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention, and FIG. 2 is a diagram showing an embodiment of a density gradation conversion table and input/output density signals. FIG. 3 is a block diagram of the second embodiment, and FIG. 4 is a density histogram. Fig. 5 is an explanatory diagram of spatial filtering processing, Fig. 6 is a diagram showing an example of a differential filter, Fig. 7 is a block diagram of the third embodiment, Fig. 8 is an explanatory diagram of smoothing processing, and Fig. 9 is a diagram showing an example of a differential filter. It is a block diagram of a 4th example. 10... Inspection object, 34... Concentration conversion means, 36... Monitor, 38. 38A-C... Table means, 40... - Setting means, 40B, 40C... Selection means, 42. ...Multi-value conversion means, 44... Defect detection means, 56... Spatial filtering means, 58... Smoothing means.

Claims (10)

[Claims] (1) In a defect inspection device that detects defects in an inspection target using a density signal obtained by scanning an image signal obtained by scanning the inspection target, a density that converts a predetermined density range that includes at least the background density into a constant density is used. table means for outputting a gradation conversion table;
A density converter converts the density signal using the density gradation conversion table; a multi-value converter converts the density signal into multiple values using the density gradation conversion table; 1. A defect inspection device comprising: defect detection means for detecting as a defect an image portion that exceeds 1000 nm and outputting a defect signal. (2) The defect inspection apparatus according to claim (1), wherein the table means is capable of manually setting a density gradation conversion table. (3) The defect inspection apparatus according to claim 1, wherein the table means obtains a density histogram of the density signal for a predetermined image area and automatically sets a density gradation conversion table from this histogram. (4) Claim (1) wherein the table means selects and outputs one of a plurality of density gradation conversion tables stored in advance.
) defect inspection equipment. (5) The defect inspection apparatus according to claim (4), wherein the table means manually selects the density gradation conversion table. (6) The defect inspection apparatus according to claim (4), wherein the table means selects one of the density gradation conversion tables using a density histogram of density signals for a predetermined image area. (7) The defect inspection apparatus according to any one of claims (1) to (6), further comprising a spatial filtering means for performing a spatial filtering process using a differential filter on the concentration signal whose concentration has been converted by the concentration converting means, and after the spatial filtering process. A defect inspection device that converts into multiple values using concentration signals. (8) The defect inspection apparatus according to claim 7, further comprising a smoothing means for smoothing the density signal after the spatial filtering process, and converting the smoothed density signal into multi-value data. (9) Claim (1) comprising a monitor that displays the density signal converted by the density converting means, so that the image of the defect can be visually viewed.
) to (8). (10) The defect inspection device according to any one of claims (1) to (8), further comprising a monitor that displays a multivalued density signal obtained by converting the density signal by the density converting means, so that an image of the defect can be visually viewed. .